Liposomal dexamethasone–moxifloxacin nanoparticle combinations with collagen/gelatin/alginate hydrogel for corneal infection treatment and wound healing

Human corneal epithelial cells (HCECs) and all the culture reagents were purchased from the American Type Culture Collection (ATCC) (Manassas, VA). The cells were maintained in fresh corneal epithelial cell basal medium (ATCC) supplemented with apo-transferrin (5 mg ml⁻¹), epinephrine (1 mM), extract P (0.4%), hydrocortisone hemisuccinate (100 ng ml⁻¹), L-glutamine (6 mM), recombinant human insulin (5 mg ml⁻¹) corneal epithelial growth factor, and 10 U ml⁻¹ penicillin/10 µg ml⁻¹ streptomycin in a CO₂ incubator at 37 °C and with 5% CO₂. The cells were subcultured by trypsinization (0.25%; Invitrogen, Carlsbad, CA) and were dissociated at approximately 80% confluence. The culture media were changed every 2–3 d.

Collagen (type I), gelatin, sodium alginate, calcium chloride, sodium citrate, 1,2-distearoyl-sn-glycero-3-phosphocholine (DSPC), cholesterol, and chloramphenicol powder were all purchased from Sigma-Aldrich (Sigma-Aldrich, St. Louis, MO).

The protocol involved in preparing liposomal antibiotics has previously been described [25]. Briefly, 50 mg of MFX combined with 7.7 mg of cholesterol and 110.6 mg of DSPC (molar ratio = 14:15:2) were dissolved in 5 ml of methylene chloride. The solvent was evaporated in a rotary evaporator at 50 °C, and the solvent traces were removed by placing the mixture in a 60 °C oven for 30 min. The phospholipid film was hydrated with either 5 or 10 mg of DEX-containing double-distilled water using a probe-type sonicator at 35 W for 30 min, which facilitated the self-assembling of Lipo-MFX/DEX. The liposomal system was then filtered using a 0.2 μM filter and was purified using a PD-10 column to deplete the non-encapsulated MFX in the supernatant.

The currently available liposomal nanoparticle-based eye drops or suspensions are not able to stay on the corneal surface for sustained drug release. To overcome this limitation, we investigated the use of a biodegradable film as a carrier to load Lipo-MFX/DEX. The biodegradable film that we used was the material that facilitated the attachment of Lipo-MFX/DEX for sustained drug release, as described previously [26]. To prepare the biodegradable film, collagen, gelatin, sodium alginate calcium chloride, and sodium citrate powder were dissolved in phosphate-buffered saline (PBS) for further experimentation, as described previously [26]. Briefly, a stock solution of gelatin and alginate was added to a 1 mg ml⁻¹ collagen solution in order to adjust the final concentration of the hydrogel to 1% alginate and 3% gelatin. Subsequently, 400 μl of the hydrogel was applied onto a 6-well cell culture plate (BD Falcon, Bedford, MA) to form a film with a thickness of ∼42 μm. The film was then chemically cross-linked by treating it with a 3% calcium chloride solution for 30 min, following which it was washed in PBS three times at room temperature.

Our first step at this stage was to establish the appropriate formulation of CGA-Lipo-MFX/DEX for ophthalmic use. Based on our degradation curve measurement, we developed a formula to determine the relationship between the molar ratio of sodium citrate to sodium alginate with the CGA decomposition time: Y = 12.3314X² − 38.6691X + 29.2180 (where Y is the CGA decomposition time, and X is the molar ratio of sodium citrate to sodium alginate) [24]. Based on this degradation formula, we designed an 8 h degradable formulation by adjusting the sodium citrate/sodium alginate ratio in our investigation.

To generate CGA-Lipo-MFX/DEX, Lipo-MFX/DEX was mixed with a liquefied CGA solution at 50 °C and was allowed to stand at room temperature for 2 h in order for the CGA complex to solidify. The solidified CGA complex was rinsed with 3% calcium chloride and was stabilized to form CGA-Lipo-MFX/DEX. The final sodium citrate concentration was calibrated and determined to be 0.01%. Owing to the properties of the hydrogel, CGA-Lipo-MFX/DEX could attach onto the corneal surface for a long time. The released MFX or DEX from the liposomes was trapped in the CGA until the matrix degraded.

TEM examinations were performed for the direct visual monitoring of the individual particle sizes of the prepared liposomes. Briefly, a drop of a sample was placed onto a carbon-coated copper grid (Polysciences Inc. Warrington PA) along with a drop of 2% phosphotungstic acid to create a thin film, and the excess solution was drained off using filter paper. The samples were also subjected to negative staining with a 2% w v^–1 aqueous uranyl acetate solution. The grid was allowed to air dry thoroughly, and the samples were then viewed using a Hitachi transmission electron microscope (Hitachi H-7650, Tokyo, Japan). The TEM images and electron diffraction patterns were visualized and collected by a soft imaging software program.

To obtain the mean size and polydispersity index (PDI), the zeta potential values were calculated using the Malvern Zetasizer Nano ZS90 system (Malvern Instruments Ltd., Malvern, UK) based on the electrophoretic mobility by means of the Helmholtz-Smoluchowski relationship [27]. The size and surface charge of the MFX and DEX loaded liposomes were measured by dynamic light scattering (DLS; Delsa™ Nano Particle Analyzer; Beckman Coulter, Fullerton, CA, USA) and zeta potential (Beckman Coulter, Brea, CA) calculations, respectively. The autocorrelation function of the scattered light was analyzed via the cumulate method, as described previously [28].

Both LC and EE are indications of the quantity of drugs entrapped within liposomal formulations. The amount of entrapped MFX was measured by spectrophotometry at a detection wavelength of 295 nm that is the maximum absorbance wavelength of MFX and was analyzed quantitatively by comparing the result with the standard curve.

To detect the quantity of DEX, samples from each group were collected and tested using a DEX enzyme-linked immunosorbent assay (ELISA) kit (MyBioSource, San Diego, CA) according to the manufacturer instructions. The plates were read at 450/630 nm dual wavelengths. The samples were tested in triplicates, and the mean values were calculated. The LC % and EE % were calculated using the following equations [29, 30]:

$$\begin{equation}{\rm{LC}}\% {\mkern 1mu} = {\mkern 1mu} \frac{{{\rm{Total}}\,{\rm{amount}}\,{\rm{of}}\,{\rm{determined}}\,{\rm{MFX}}\,{\rm{or}}\,{\rm{DEX}}\,{\rm{in}}\,{\rm{the}}\,{\rm{liposome}}}}{{{\rm{Total}}\,{\rm{amount}}\,{\rm{ofMFX}}\,{\rm{or}}\,{\rm{DEX}}\,{\rm{liposome}}}} \times 100\end{equation} \tag{ 1 }$$

$$\begin{array}{*{20}{c}} {{\rm{EE}}\% = \frac{{{\rm{Total}}\,{\rm{amount}}\,{\rm{of}}\,{\rm{determined}}\,{\rm{MFX}}\,{\rm{or}}\,{\rm{DEX}}\,{\rm{in}}\,{\rm{the}}\,{\rm{liposome}}}}{{{\rm{Initial}}\,{\rm{amount}}\,{\rm{of}}\,{\rm{MFX}}\,{\rm{or}}\,{\rm{DEX}}\,{\rm{loading}}}} \times 100} \end{array} \tag{ 2 }$$

Based on our previous investigation, we established the formula to determine the relationship between the molar ratio of sodium citrate to sodium alginate with CGA decomposition time [24]. The CGA complex completely degraded in 0.01% sodium citrate. The in vitro release profiles of MFX or DEX from CGA-Lipo-MFX/DEX were studied, as described previously [31]. Briefly, a 2 ml volume of CGA-Lipo-MFX/DEX was put in a dialysis bag (molecular weight cutoff 3.5 kDa) with 0.01% sodium citrate, and the dialysis bag was incubated in 50 ml of PBS (pH ∼ 7.4) and gently stirred at 37 °C, and the release medium was collected at predetermined time intervals. To detect the amount of released MFX, the samples were quantified via spectrophotometry at a detection wavelength of 295 nm and were analyzed quantitatively by comparisons with the standard curve, as described above.

To detect the amount of released DEX, the samples were quantified using a DEX ELISA kit (MyBioSource, San Diego, CA) according to the manufacturer's instructions, as described above. The samples were tested in triplicates, and the mean values were calculated. All the results were expressed as mean ± standard deviation.

Bacillus strains (Escherichia coli (E. coli)) were cultured in sterilized glass flasks containing 250 ml of BG-11 medium, as described previously [32]. The temperature was maintained at 37 °C in a standard incubator. No bacterial contamination was detected during a microscopic examination of the culture. The growth curve of E. coli was determined by spectrophotometry at a wavelength of 600 nm.

To determine the growth inhibition of E. coli induced by CGA- Lipo-MFX/DEX, 50 ml E. coli suspensions were treated with the control, CGA only, or the following formulations: CGA-Lipo-MFX, CGA-Lipo-DEX, or CGA-Lipo-MFX/DEX. E. coli suspensions were collected at predetermined time intervals, and the concentration was measured by spectrophotometry, as described above. The inhibition efficiency was calculated using the following standard formula:

$$\begin{align*}& E.\ coli\ \text{inhibition efficiency (%)}\\ &\quad= \text{[(control - treatment)/control]}*100\end{align*}$$

To further investigate the growth inhibition effects induced by CGA- Lipo-MFX/DEX on E. coli, the direct colony counting method was also performed, as described previously [33]. Briefly, 10 μl E. coli suspensions collected from the samples with the various slow antibiotic-releasing formulations at different time intervals were added to 500 μl of BG-11 medium and then cultured on a BG-11 agar plate in an incubator at 37 °C for 12 h. The colony-counting numbers were calibrated using the BioCapt v.11.02 software program (Vilber Lourmat, Cedex, France).

To evaluate and quantify corneal epithelial cell cytotoxicity resulting from CGA-Lipo-MFX/DEX, a traditional direct cell counting method was performed, as previously described, with some modifications [34, 35]. Briefly, CGA, CGA-Lipo-MFX, CGA-Lipo-DEX, and CGA-Lipo-MFX/DEX were first placed in 6-well plates, onto which 1 × 10⁵ corneal epithelial cells were seeded. The cells were then incubated in complete corneal epithelial cell growth medium for 12 h. After incubation, the cells were trypsinized, and cell suspensions in 1 × PBS were incubated with trypan blue (0.4%) (Sigma-Aldrich, St. Louis, MO) (1:1 volume) for 5 min and were assessed using a hemocytometer. The cell numbers in each sample were counted in five random experiments, and the percentage of cytotoxic cells was calculated as the number of CGA-Lipo-MFX, CGA-Lipo-DEX, or CGA-Lipo-MFX/DEX cells/the number of CGA-treated cells that were counted.

To evaluate and quantify the wound healing effect of CGA-Lipo-MFX/DEX in vivo, the CGA-Lipo-MFX/DEX treatment was administered to a central corneal epithelial debridement mouse model, as described previously [36, 37]. Briefly, C57BL/6 mice were anesthetized via intraperitoneal injections of ketamine hydrochloride (2 mg g⁻¹ body weight) and xylazine (0.4 mg g⁻¹ body weight). After the topical application of one drop of proparacaine (Alcaine; Alcon Laboratories, Inc. Fort Worth, TX) in each eye, the central cornea was marked using a trephine with a diameter of 2 mm, and the epithelium was debrided using a corneal rust ring remover with a 0.5 mm burr (Algerbrush IITM; Alger Equipment Co., Inc. Lago Vista, TX) under a stereomicroscope (SV11; Carl Zeiss Meditec, Dublin, CA). The animals were sacrificed, and the eyeballs were isolated and then cultured in a 96-well plate. CGA, CGA-Lipo-MFX, and CGA-Lipo-MFX/DEX were seeded onto the top of the eyeballs. Subsequently, the specimens were cultured in sodium citrate-containing Dulbecco's modified eagle medium (DMEM) (Invitrogen-Gibco, Grand Island, NY) containing 1% fetal bovine serum (FBS) without antibiotics in a humidified atmosphere of 5% CO₂ at 37 °C. After incubation, the extent of corneal wound closure was examined by fluorescein staining and was photographed with a digital camera (Axiovision; Carl Zeiss Meditec GmbH, Oberkochen, Germany). The eyeballs were then fixed in 4% paraformaldehyde in PBS and were embedded in paraffin for further experimentation.

The epithelium-debrided eyeballs from the mice were isolated and treated with CGA, CGA-Lipo-MFX, or CGA-Lipo-MFX/DEX and were cultured as described above. The specimens were then fixed in 10% buffered neutral formalin and were embedded in paraffin. Sections were cut at a thickness of 3–5 μm and were stained with hematoxylin and eosin. The histopathological changes, including cell morphology and the presence of metastatic tumor cells, were examined by light microscopy (40X and 200X).

Statistical analyses were carried out using SPSS version 6.0 for Windows and involved a one-way analysis of variance (ANOVA) and the Mann–Whitney U test. A P value of <0.05 was considered to be statistically significant.

We synthesized Lipo-MFX/DEX at various doses based on different ratios of MFX/DEX, including 50:5 and 50:10, as described previously. The Lipo-MFX/DEX nanoparticle was observed by TEM (figure 1(A)). The TEM images of the Lipo-MFX/DEX morphology indicated that the sizes and spherical shapes of these nanoparticles were consistent with those revealed in the DLS analysis (figure 1(B) and table 1). The sizes of the Lipo-MFX/DEX particles with the ratios of 50:5 and 50:10 were 213.4 ± 171.8 nm and 435.3 ± 316.6 nm, respectively (figure 1(C) and table 1). The majority of the particles were approximately 0.2 nm in size, demonstrating that the solutions were homogeneous with a narrow size distribution. In further experiments, the EE of the designed MFX/DEX-loaded nanoparticles was examined. As shown in table 1, the EE of Lipo-MFX/DEX was 94.5 ± 3.8% (50:5) and 84.6 ± 5.2% (50:10), the drug content was 18.1 ± 1.4% (50:5) and 12.4 ± 2.2% (50:10), and the zeta potential was −4.8 ± 3.3 mV (50:5) and −3.7 ± 3.1 mV (50:10) (table 1).

Zoom In Zoom Out Reset image size

These results indicate that all the designed nanoparticles were successfully prepared at the nanoscale level; however, in comparison with the 50 mg MFX combined with 10 mg DEX-loaded liposomal nanoparticles, the 50 mg MFX combined with 5 mg DEX-loaded liposomal nanoparticles were superior in terms of EE and drug content. Consequently, we used the 50 mg MFX combined with 5 mg DEX-loaded liposomal nanoparticles as the standard preparation for further studies.

Next, we investigated the release profiles of MFX and DEX from Lipo-MFX/DEX in vitro. The release profile of MFX from the Lipo-MFX/DEX nanoparticles was evaluated by the dialysis bag diffusion method in PBS at 37 °C. It took 10 min for the Lipo-MFX/DEX formulation to release an effective dose of MFX (MIC₉₀ = 1.5 μg ml⁻¹). The following are the correlations between the timepoints and doses of MFX released: control, −0.01 ± 0.0 μg ml⁻¹; 10 min, 2.2 ± 0.3 μg ml⁻¹; 20 min, 3.9 ± 0.3 μg ml⁻¹; 30 min, 5.8 ± 0.4 μg ml⁻¹; 45 min, 8.2 ± 0.5 μg ml⁻¹; 60 min, 10.0 ± 1.2 μg ml⁻¹; 120 min, 16.3 ± 1.4 μg ml⁻¹; 240 min, 19.9 ± 1.4 μg ml⁻¹; 360 min, 21.3 ± 2.0 μg ml⁻¹; 480 min, 20.1 ± 2.3 μg ml⁻¹; 720 min, 20.1 ± 1.7 μg ml⁻¹; and 1440 min, 19.8 ± 2.0 μg ml⁻¹ (figure 2(A)).

Zoom In Zoom Out Reset image size

The time-based DEX release profile of Lipo-MFX/DEX was also measured, as described. It took 30 min for Lipo-MFX/DEX to release an effective dose of DEX (40 ng ml⁻¹). The correlations between the timepoints and doses of DEX released were similar to those of MFX (from −0.00 ± 0.00 to 4.9 ± 0.5 μg ml⁻¹ at each timepoint, respectively, figure 2(B)).

These results reveal that the MFX and DEX encapsulated liposomal nanoparticles could release increasing concentrations of both the drugs. Not only could the Lipo-MFX/DEX complex release the effective dose/concentration of MFX and DEX in a short time, but it could also sustain the MFX and DEX release for a long period.

We established the validity of the CGA complex degradation formula in our previous investigation. We also confirmed the EE and drug release profiles of Lipo-MFX/DEX. To develop a new effective MFX/DEX formulation for ophthalmic application, we combined the CGA complex with Lipo-MFX/DEX to form CGA-Lipo-MFX/DEX, a biomatrix loaded with liposomal MFX and DEX, which was formulated to degrade in 8 h. We then investigated the release of both MFX and DEX from CGA-Lipo-MFX/DEX. Figure 2 illustrates the results of the investigation of MFX and DEX-release from CGA-Lipo-MFX/DEX. As shown in figure 2, the release of MFX (figure 2(C)) and DEX (figure 2(D)) from CGA-Lipo-MFX/DEX resulted in an effective dose in 30 min (The timepoint-drug release correlation ranged from −0.0 ± 0.0 to 19.3 ± 2.4 μg ml⁻¹ for MFX, figure 2(C)) and 60 min (The timepoint-drug release correlation ranged from −0.0 ± 0.0 to 3.5 ± 0.4 μg ml⁻¹ at each timepoint for DEX, figure 2(D)), respectively.

Based on the results presented in figures 2(C) and (D), we confirmed that the release of both MFX and DEX from CGA-Lipo-MFX/DEX could be sustained stably. This formulation could not only generate an effective dose/concentration of MFX and DEX in a short time but could also sustain the release of MFX and DEX for a long period.

We previously evaluated the release profiles of both MFX and DEX from CGA-Lipo-MFX/DEX in vitro. CGA-Lipo-MFX/DEX could generate an effective dose/concentration in 30 min and sustain the release of MFX and DEX for more than 12 h. We then investigated the efficacy of various CGA loaded MFX/DEX liposomal nanoparticles in inhibiting bacterial growth activity. An E. coli growth model was used to investigate MFX-inhibited pathogen proliferation in the current study. In the experiment, E. coli with the addition of normal saline at specific timepoints was used as the control group (100%). As shown in figure 4, CGA-Lipo-MFX/DEX could stably and constitutively suppress E. coli proliferation, which started in 2 h and was sustained for more than 24 h. The viability rate of E. coli at each timepoint was as follows: 2 h, 67.4 ± 15.7%; 4 h, 48.5 ± 10.2%; 6 h, 39.7 ± 5.9%; 8 h, 28.6 ± 4.2%; 12 h, 17.1 ± 3.3%; and 24 h, 17.4 ± 5.1% (figure 3(A)).

Zoom In Zoom Out Reset image size

To quantitate the inhibitory effect induced by CGA-Lipo-MFX/DEX on E. coli growth, direct colony counting experiments were conducted. The representative bacterial growth status at various timepoints is presented in figure 3(B). As shown in figure 3(C), CGA-Lipo-MFX/DEX significantly inhibited the number of E. coli colonies, which also started at the 2 h timepoint and was sustained for 24 h. The maximum inhibition effect was approximately 55.1% on allowing CGA-Lipo-MFX/DEX to act on the E. coli suspensions for 6 h.

The aforementioned evidence indicates that CGA-Lipo-MFX/DEX could not only generate an effective dose/concentration of both MFX and DEX in a short time and sustain MFX or DEX release for a long period, but it could also inhibit E. coli proliferation.

Our previous experiments have revealed the properties of CGA-Lipo-MFX/DEX in terms of rapidly generating an effective dose/concentration, sustaining the stable release of MFX/DEX, and inhibiting E. coli growth for a long period. We also investigated the biocompatibility of CGA-Lipo-MFX/DEX and whether this novel ophthalmic formulation produced corneal epithelial toxicity. Consequently, in the current study, we co-cultured corneal epithelial cells with CGA, CGA-Lipo-MFX, CGA-Lipo-DEX, or CGA-Lipo-MFX/DEX and then calibrated the viable cells in order to analyze the potential toxicity. Cells counts from the co-culture with CGA for 24 h were defined as the control group results (100%, figure 5). As shown in figure 4, the cell survival rate in the CGA-Lipo-DEX- and CGA-Lipo-MFX/DEX-treated groups were significantly higher compared to in the other groups (CGA-Lipo-MFX: 91.5 ± 3.5%, CGA-Lipo-DEX: 186.7 ± 18.1%, and CGA-Lipo-MFX/DEX: 216.2 ± 20.7, figure 4). No significant cell toxicity was noted in any of the groups. This evidence suggests that all the sustained MFX/DEX-releasing formulations including CGA-Lipo-MFX, CGA-Lipo-DEX, and CGA-Lipo-MFX/DEX did not induce significant toxicity in corneal epithelial cells. Moreover, with the addition of DEX, the epithelial cell proliferation increased. The proliferated cells might help in promoting corneal epithelial wound healing.

Zoom In Zoom Out Reset image size

The previous evidence that we have presented has demonstrated that CGA-Lipo-MFX/DEX could not only suppress pathogenic microorganism growth but could also increase corneal epithelial cell proliferation. To validate these findings in vivo, we topically administered exogenous CGA-Lipo-MFX/DEX onto cultured corneal epithelial debridement mouse eyes in order to observe the anti-infection and corneal re-epithelialization effects. As shown in figure 5, the corneal epithelial cells were covered with CGA, CGA-Lipo-MFX, or CGA-Lipo-MFX/DEX on day 0. The eyeball became turbid in the CGA-treated group on day 1. The eyeballs in the CGA-Lipo-MFX or CGA-Lipo-MFX/DEX-treated groups remained clear until day 3. These results indicate that CGA-Lipo-MFX and CGA-Lipo-MFX/DEX could significantly inhibit microorganism growth in vivo.

Zoom In Zoom Out Reset image size

We subsequently determined if there were any differences in the levels of infiltration of leukocytes or lymphocytes within in the different eyeball samples treated with the various CGA loaded MFX/DEX liposomal nanoparticles. The corneal epithelial debridement mice were sacrificed, the eyeballs were excised, and tissue blocks were prepared using the CGA-, CGA-Lipo-MFX-, or CGA-Lipo-MFX/DEX-treated eyeballs, as described earlier. Representative figures of the infiltrating leukocytes within the eyeballs of the different treatment groups evaluated by hematoxylin and eosin staining are presented in figure 6. The corneal epithelial debridement eyeballs that were treated with CGA only exhibited increased inflammation and also edema progression without epithelial regeneration. As shown in figure 6, the corneal epithelial cells were scraped on day 1 in all groups. Epithelial cell growth was not enhanced in the mice that received only CGA as observed on day 3. In comparison with the CGA-treated group, the CGA-Lipo-MFX-treated group and especially the CGA-Lipo-MFX/DEX-treated group exhibited not only significantly decreased edema and inflammation but also enhanced epithelial cell regeneration. The number of infiltrating leukocytes was also dramatically reduced in the eyeballs treated with CGA-Lipo-MFX and especially in those treated with CGA-Lipo-MFX/DEX. This evidence indicates that the CGA-Lipo-MFX/DEX formulation produces dual effects including anti-bacterial growth and anti-inflammatory effects, which enhance corneal epithelial cell proliferation, thereby accelerating corneal wound healing.

Zoom In Zoom Out Reset image size